ardzdirect
This model is a fine-tuned version of facebook/mms-1b-all on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2738
- Wer: 0.4396
- Bleu: 0.3030
- Rouge: {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0}
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer | Bleu | Rouge |
---|---|---|---|---|---|---|
2.9801 | 0.8316 | 100 | 0.4689 | 0.6744 | 0.0952 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.5487 | 1.6570 | 200 | 0.3861 | 0.6112 | 0.1243 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4999 | 2.4823 | 300 | 0.3645 | 0.5976 | 0.1358 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.482 | 3.3077 | 400 | 0.3483 | 0.5838 | 0.1788 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4421 | 4.1331 | 500 | 0.3513 | 0.5801 | 0.1489 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4465 | 4.9647 | 600 | 0.3361 | 0.5699 | 0.1562 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4425 | 5.7900 | 700 | 0.3482 | 0.5886 | 0.1368 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4049 | 6.6154 | 800 | 0.3304 | 0.5537 | 0.1708 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4011 | 7.4407 | 900 | 0.3155 | 0.5451 | 0.2087 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.4033 | 8.2661 | 1000 | 0.3319 | 0.5522 | 0.1835 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3804 | 9.0915 | 1100 | 0.2983 | 0.5172 | 0.2160 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3672 | 9.9231 | 1200 | 0.2957 | 0.5170 | 0.2282 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3661 | 10.7484 | 1300 | 0.2999 | 0.5152 | 0.2346 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3599 | 11.5738 | 1400 | 0.2952 | 0.4998 | 0.2372 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3475 | 12.3992 | 1500 | 0.3041 | 0.4961 | 0.2493 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3387 | 13.2245 | 1600 | 0.2971 | 0.5205 | 0.2432 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3474 | 14.0499 | 1700 | 0.2948 | 0.4808 | 0.2616 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3238 | 14.8815 | 1800 | 0.2809 | 0.4860 | 0.2669 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3156 | 15.7069 | 1900 | 0.2740 | 0.4544 | 0.2857 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3185 | 16.5322 | 2000 | 0.2736 | 0.4627 | 0.2799 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3148 | 17.3576 | 2100 | 0.2793 | 0.4455 | 0.2959 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.2996 | 18.1830 | 2200 | 0.2705 | 0.4451 | 0.2943 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.3112 | 19.0083 | 2300 | 0.2708 | 0.4440 | 0.2986 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
0.2916 | 19.8399 | 2400 | 0.2738 | 0.4396 | 0.3030 | {'rouge1': 0.0, 'rouge2': 0.0, 'rougeL': 0.0, 'rougeLsum': 0.0} |
Framework versions
- Transformers 4.49.0
- Pytorch 2.6.0+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0
- Downloads last month
- 7
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
๐
Ask for provider support
Model tree for ilyes25/ardzdirect
Base model
facebook/mms-1b-all